Effect of organic macerals on hydrocarbon generation ability of oil shale and coal and establishment of a contribution model – constraints from machine learning

To quantify the impact of organic macerals on hydrocarbon generation in oil shale and coal, this study analyzed 403 datasets. The results show distinct differences in maceral composition between the two materials. Using a random forest model with aquatic liptinite, terrigenous liptinite, vitrinite, and inertinite as inputs, TOC and HI predictions achieved R² values of 0.62 and 0.73, respectively. SHAP analysis reveals that aquatic liptinite is the dominant positive contributor to both parameters, while terrigenous liptinite shows a bidirectional effect. Based on these relationships, a quantitative contribution model for organic macerals in oil shale is established, offering insights for resource evaluation and exploration.